测量不确定度估计的极限费舍尔信息方法

Extreme Fisher Information Approach for Measurement Uncertainty Evaluation

极限费舍尔信息(EFI)是源于极限物理信息理论下的一种信息测度。由于在测量实践中,很难一一准确且高效地定义与补偿所有影响测量结果的因素并估计测量不确定度。因此,该文提出了采用根据EFI推导的概率密度函数(PDFs)来估计被测量的测试边界信息,即待测系统的测量不确定度。该方法能够根据不同的不确定度影响因素以及待测系统的物理规则更加动态地刻画测量不确定性。从物理应用角度进行了详细的数理推导与讨论,相比不考虑物理意义的数学模型,该方法更适用于实际应用。最后,用两组实例验证了该EFI方法的有效性。

The extreme Fisher information （EFI） is originally a measure within the theory of extreme physical information （EPI）. In measurement activities, it is hard to accurately and efficiently identify and compensate every effect in measurement and evaluate the incompleteness of the measurement results. So we propose to employ the probability density functions （PDFs） derived from the EFI for estimating the boundary information of the measurement results, that is, the associated measurement uncertainty. The proposed method can characterize the measurement uncertainty more dynamically, with considering the different behaviors of the uncertainty effects and the law governing the system under measurement at the same time. The proposed approach yields the possible distribution of the measurement result in a more practical way rather than the pure mathematical approach, which is more applicable. Finally, the effectiveness of the proposed EFI method is demonstrated by the numerical results of two practical instances.